Richard A. Buswell

Identification of evolving fuzzy rule-based models

An approach to identification of evolving fuzzy rule-based (eR) models is proposed. eR models implement a method for the noniterative update of both the rule-base structure and parameters by incremental unsupervised learning. The rule-base evolves by adding more informative rules than those that previously formed the model. In addition, existing rules can be replaced with new rules based on ranking using the informative potential of the data. In this way, the rule-base structure is inherited and updated when new informative data become available, rather than being completely retrained. The adaptive nature of these evolving rule-based models, in combination with the highly transparent and compact form of fuzzy rules, makes them a promising candidate for modeling and control of complex processes, competitive to neural networks. The approach has been tested on a benchmark problem and on an air-conditioning component modeling application using data from an installation serving a real building. The results illustrate the viability and efficiency of the approach.

Automatic generation of fuzzy rule-based models from data by genetic algorithms

A methodology for the encoding of the chromosome of a genetic algorithm (GA) is described in the paper. The encoding procedure is applied to the problem of automatically generating fuzzy rule-based models from data. Models generated by this approach have much of the flexibility of black-box methods, such as neural networks. In addition, they implicitly express information about the process being modelled through the linguistic terms associated with the rules. They can be applied to problems that are too complex to model in a first principles sense and can reduce the computational overhead when compared to established first principles based models. The encoding mechanism allows the rule base structure and parameters of the fuzzy model to be estimated simultaneously from data. The principle advantage is the preservation of the linguistic concept without the need to consider the entire rule base. The GA searches for the optimum solution given a comparatively small number of rules compared to all possible. This minimises the computational demand of the model generation and allows problems with realistic dimensions to be considered. A further feature is that the rules are extracted from the data without the need to establish any information about the model structure a priori. The implementation of the algorithm is described and the approach is applied to the modelling of components of heating ventilating and air-conditioning systems.

Demonstration of Fault Detection and Diagnosis Methods for Air-Handling Units

Results are presented from controlled field tests of two methods for detecting and diagnosing faults in HVAC equipment. The tests were conducted in a unique research building that featured two air-handling units serving matched sets of unoccupied rooms with adjustable internal loads. Tests were also conducted in the same building on a third air handler serving areas used for instruction and by building staff. One of the two fault detection and diagnosis (FDD) methods used first-principles-based models of system components. The data used by this approach were obtained from sensors typically installed for control purposes. The second method was based on semiempirical correlations of submetered electrical power with flow rates or process control signals. Faults were introduced into the air-mixing, filter-coil, and fan sections of each of the three air-handling units. In the matched air-handling units, faults were implemented over three blind test periods (summer, winter, and spring operating conditions). In each test period, the precise timing of the implementation of the fault conditions was unknown to the researchers. The faults were, however, selected from an agreed set of conditions and magnitudes, established for each season. This was necessary to ensure that at least some magnitudes of the faults could be detected by the FDD methods during the limited test period. Six faults were used for a single summer test period involving the third air-handling unit. These fault conditions were completely unknown to the researchers and the test period was truly blind. The two FDD methods were evaluated on the basis of their sensitivity, robustness, the number of sensors required, and ease of implementation. Both methods detected nearly all of the faults in the two matched air-handling units but fewer of the unknown faults in the third air-handling unit. Fault diagnosis was more difficult than detection. The first-principles-based method misdiagnosed several faults. The electrical power correlation method demonstrated greater success in diagnosis, although the limited number of faults addressed in the tests contributed to this success. The first-principles-based models require a larger number of sensors than the electrical power correlation models, although the latter method requires power meters that are not typically installed. The first-principles-based models require training data for each subsystem model to tune the respective parameters so that the model predictions more precisely represent the target system. This is obtained by an open-loop test procedure. The electrical power correlation method uses polynomial models generated from data collected from “normal” system operation, under closed-loop control. Both methods were found to require further work in three principal areas: to reduce the number of parameters to be identified; to assess the impact of less expensive or fewer sensors; and to further automate their implementation. The first-principles-based models also require further work to improve the robustness of predictions.

Evolving rule-based models: A tool for intelligent adaptation

An online approach for rule-base evolution by recursive adaptation of rule structure and parameters is described . An integral part of the procedure is to maximise the model transparency by simplifying the fuzzy linguistic descriptions of the input variables. The rule base evolves over time, utilising direct calculation approaches and hence minimising the reliance on the use of computationally expensive techniques, such as genetic algorithms. An online version of subtractive clustering recently introduced by the authors (P.P. Angelov and R.A. Buswell) is used for determination of the antecedent part of the fuzzy rules. Recursive least squares estimation is employed to determine the parameters of the consequent part of each rule. The use of these efficient non-iterative techniques is the key to the low computational demands of the algorithm. The application of similarity measures improves the interpretability and compactness of the resulting eR model, with no significant detriment to the model precision. A time series prediction problem on data from a real indoor climate control (ICC) system has been considered to test and validate the proposed model simplification method.

Development of a Viable Concrete Printing Process

A novel Concrete Printing process has been developed, inspired and informed by advances in 3D printing, which has the potential to produce highly customised building components. Whilst still in their infancy, these technologies could create a new era of architecture that is better adapted to the environment and integrated with engineering function. This paper describes the development of a viable concrete printing process with a practical example in designing and manufacturing a concrete component (called Wonder Bench) that includes service voids and reinforcement. The challenges met and those still to be overcome particularly in the evaluation of the manufacturing tolerances of prints are also discussed.

Evolutionary Synthesis of HVAC System Configurations: Algorithm Development (RP-1049)

This paper describes the development of a model-based optimization procedure for the synthesis of novel heating, ventilating, and air-conditioning system configurations. The optimization problem can be considered as having three suboptimization problems: the choice of a component set; the design of the topological connections between the components; and the design of a system operating strategy. In an attempt to limit the computational effort required to obtain a design solution, the approach adopted in this research is to solve all three subproblems simultaneously. The computational effort has been further limited by implementing simplified component models and including the system performance evaluation as part of the optimization problem (there being no need, in this respect, to simulate the system performance). The optimization problem has been solved using a Genetic Algorithm (GA) that has data structures and search operators specifically developed for the solution of HVAC system optimization problems. The performance of the algorithm and various search operators has been examined for a two-zone optimization problem, the objective of the optimization being to find a system design that minimizes system energy use. In particular, the performance of the algorithm in finding feasible system designs has been examined. It was concluded that the search was unreliable when the component set was optimized, but if the component set was fixed as a boundary condition on the search, then the algorithm had an 81% probability of finding a feasible system design. The optimality of the solutions is not examined in this paper but is described in an associated publication (Wright and Zhang 2008). It was concluded that, given a candidate set of system components, the algorithm described here provides an effective tool for exploring the design of novel HVAC systems.

Automatic design synthesis and optimization of component-based systems by evolutionary algorithms

A novel approach for automatic design synthesis and optimization using evolutionary algorithms (EA) is introduced in the paper. The approach applies to component-based systems in general and is demonstrated with a heating, ventilating and air-conditioning (HVAC) systems problem. The whole process of the system design, including the initial stages that usually entail significant human involvement, is treated as a constraint satisfaction problem. The formulation of the optimization process realizes the complex nature of the design problem using different types of variables (real and integer) that represent both the physical and the topological properties of the system; the objective is to design a feasible and efficient system. New evolutionary operators tailored to the component-based, spatially distributed system design problem have been developed. The process of design has been fully automated. Interactive supervision of the optimization process by a human-designer is possible using a specialized GUI. An example of automatic design of HVAC system for two-zone buildings is presented.

Benchmarking small power energy consumption in office buildings in the United Kingdom: A review of data published in CIBSE Guide F

CIBSE’s Guide F is a widely recognised guidance document on energy efficiency in buildings, which includes energy consumption benchmarks for small power equipment in offices. In its recently published 3rd edition, existing power demand benchmarks for office equipment were revised to better represent appliances found in contemporary office buildings. Other key sources of data such as typical operating hours for equipment, however, have been omitted. This paper compares the benchmarks published in both the 2nd and 3rd editions of Guide F against a set of measurements of small power loads in a real UK office building. Load profiles for the monitored equipment are also presented to supplement the information included in the new Guide F. Practical application: With the increasing demand for more realistic predictions of operational energy use in buildings, small power should not be disregarded since it typically accounts for more than 20% of total energy used in offices. Furthermore, small power loads can have a significant impact on the cooling loads of a building. This paper reviews existing benchmarks, focusing on the new update to CIBSE Guide F, comparing available benchmarks against newly gathered monitored data. Detailed load profiles for individual office equipment are also provided, which can be used by designers to inform better predictions of small power consumption in office buildings.

Uncertainty in model-based condition monitoring

Model-based techniques for automated condition monitoring of HVAC systems have been under development for some years. The generation of false alarms has been identified as a principal factor affecting the potential usefulness of condition monitoring in HVAC applications. Results from the application of these methods to systems installed in real buildings have highlighted the difficulty in selecting good alarm thresholds that balance robustness (lack of false alarms) and sensitivity (early detection). This paper demonstrates that this balance can be met in a transparent and analytical manner, through the application of uncertainty analysis. The paper discusses the sources of uncertainty associated with component models and system measurements. A condition monitoring scheme applied to a typical HVAC cooling coil subsystem installed in a real building is presented. Faults are artificially introduced into the system and are used in conjunction with fault-free operation to demonstrate the sensitivity and robustness of the scheme. The principle conclusions drawn by the paper consider the likely minimum magnitudes of faults that can be detected in typical HVAC systems, without false alarm generation. More broadly however, the paper demonstrates that the issue of uncertainty affects all aspects of system monitoring, modelling and control.

Multidisciplinary research: should effort be the measure of success?

ABSTRACT Energy demand reduction and flexible demand from dwellings will play a critical role in achieving a low-carbon future. There remain many unanswered questions around the interaction of people with their environment and the technical systems that service them and, as a result, multidisciplinary research is a principal component of research funding internationally. However, relatively little published work considers the operational issues in undertaking epistemologically diverse, academic research projects. This paper makes a contribution by quantifying the operational effort involved in data collection on a large multidisciplinary project and connecting the operational issues encountered to knowledge production. It is found that the cost of the data gathering is £46,000/home, and participants can give upwards of 217 hours of their time per house engaging with data-gathering activities. The rate of knowledge production is found to be approximately three publication/full-time equivalents (FTE) over the lifetime of the project and the risk to generating interdisciplinary insights is shown to be dependent on largely unforeseeable operational issues that compound the characteristic differences in the collection of the data utilized by social and technical research communities.